Oracle 18c and Its Impact on DBAs

Oracle Database 18c the next generation of the industry-leading database.

Oracle introduces the world's first autonomous database cloud. Oracle Autonomous Database Cloud eliminates complexity, human error, and manual management, helping to ensure higher reliability, security, and more operational efficiency at the lowest cost.

Copied well written article by Joel Per├ęz and Skant Gupta regarding Oracle 18c from Oracle community which will definitely help to DBAs who are worrying about future. Below is the full article.

“DBAs will be fired.. ? It’s the end of DBAs because of the Autonomous Databases..?

Obviously not.. believe it or not.. now the DBAs will have more important role in enterprises.

I have +17 years as Oracle DBA and I could tell you how this profession have been evolving, having a tendency of:
  • More elaborated and advanced features.. that’s the essence of every new release..
  • More automated tasks release by release.. the DBAs and Enterprises must realise real advantages to perform early or convenient upgrades before the “Non support of Release” force them to do it..
  • and above all.. how to become DBA each day is even more complex than before.. and who will be affected by this.. this is the main point I want to focus in this short article.

Oracle Database 18c the next generation of the industry-leading database.
Oracle introduces the world's first autonomous database cloud. Oracle Autonomous Database Cloud eliminates complexity, human error, and manual management, helping to ensure higher reliability, security, and more operational efficiency at the lowest cost.

At the previous paragraph I have highlighted the sentences "database cloud" & "Oracle Autonomous Database Cloud", it means.. when we are talking about Cloud Databases, Oracle Autonomous Database Cloud is just another different kind of Cloud database service, so, in this article we will call it simply as "Cloud Databases".

Autonomous Databases/Cloud Databases.. is a subject that could be analyzed from different points of views. Most of articles and publications focus in how ideal would be a futuristic database etc..

but Let’s think for a minute..
Who will decide under what kind of service the databases will be.. ?
Who will implant the policies for those autonomous databases ?
Who will have the knowledge to decide or estimate the cost of these services based on the regular activities of those databases ?
Who will decide the complex design of IT Infrastructure when we have more and more options.. ?
The answer is obvious.. a DBA.. not any simple DBA.. a DBA that has evolved with all this new generation of databases on Cloud.

Oracle Autonomous Database as I mentioned before is a kind of different type of a Cloud Database. So.. in order to move that database from "On-prem" or create it from zero, a DBA must be needed..

Oracle Autonomous Database is a kind of database with a lot of mundane tasks already automated such as:
  • Applying of Patches
  • Upgrades
  • and Tunes itself..

The more important section of this article is the following

I will use the exact smart words of my colleague Tim Hall ( Oracle ACE Director ) "18c on-prem will have very little impact on DBAs. It is the Autonomous Database suite of services that have the impact"
  • “18c on-prem will have very little impact on DBAs”, Yes.. Exactly, Oracle 18c is just another higher version, “Oracle 18c” is not a relational database manager software that will work by default in autonomus mode
  • “It is the Autonomous Database suite of services that have the impact”, Autonomous databases is a kind of suit of services that will be available for “Oracle Public Cloud”, this means that “On-Prem” databases will not work in autonomous mode “until what we know up to now..”, perhaps later that kind of service could be available for “On-prem databases”

Most of DBAs currently are in fear believing “Autonomous Databases suite of Services” will let them without job.. but the reality is not even near to that.. actually there are several kind of services to lodge a database on Cloud:
  • Oracle Database Cloud Service
  • Oracle Bare Metal Cloud Database Service
  • Oracle Database Exadata Cloud Service
  • Oracle Database Exadata Cloud Machine
  • Oracle Database Express Cloud Service
“Autonomous Databases suite of Services”  will represent just other possible services you could contract
Oracle 18c is not Autonomous Database suit of Services
Autonomous Database Service suit of Services is not Oracle 18c
Autonomous Database suit of Services is a kind of services available up to now just for Oracle Public Cloud
Oracle 18c is just one release more

So.. with those lines you have got the more important section of this article.
The conclusion is that "Oracle 18c" will have little impact for the regular "On-prem" databases.

Now, Let's extend this conversation..

But Let's wait a minute.. Let's analyze every one of those.. how it can be done today with the regular "Oracle Database Cloud Service (DBCS)" and how we think Oracle will do it.. We talk about "How Oracle will do it.. because the debut of Autonomous Datawarehouse Database Cloud will be in December 2017.

Applying of Patches

Nowadays, if you want to apply a Patch using DBCA is just to go to the console, see on the screen if there is any available patch for that database and make some clicks to apply it.. So simple..

So.. I think is not "so difficult" for Oracle Corp, replace our clicks for an automated process..

There are some patches that require the database must be stopped because they change binaries.. etc.. well.. most probable Oracle had already a mechanism in order to apply patches having the database running and they have decided to release it now.. 

If we imagine how work the database inside running sequentially sentences.. etc.. every sentence.. operations.. etc all this can be coordinated so that patches affect in some way "post" sentences.. etc.. in a general way what I'm trying to say is.. the fact of applying patches by itself is not the big thing taking into account the level of development of the Oracle product during years.. So.. the applying patch by itself is just a step forward related to levels of automation.


Up to now when you are working with DBCS the only way to upgrade a database already working in the cloud is creating another service which will have another compute node and we apply a regular procedure to upgrade that database. However, we have to remember that Oracle has been working very hard in establish very advance mechanism to manage PDB. Nowadays we can clone PDB in hot, we can move PDBs from one container to another container in hot.. basically is very alike to the mechanism implanted to move a datafile online.. so.. this technology is extreme mature by the Oracle Software in these times.. So.. 

Upgrade a database must be surely a procedure a alike to what is done with PDB, the CDB in its root definition has the binaries of the corresponding version and we move just the data. I'm "sure".. that's the mechanism the Oracle software will use to upgrade databases in hot. So.. if we realise all I have commented.. all this is just a technology that we are using with the current release.. difference in autonomous databases is that they are applying this procedure with complete databases under this kind of new service.

Note : later of writing this article my colleague Tim Hall ( Oracle ACE Director ) had helped me to clarify that element, these are the exact words of Tim
"Oracle have explained how the upgrades and patches will work. These systems are based on 18c running on Exadata. Since 18c allows rolling upgrades and patches for everything, including OJVM, they are able to do online patches for these services"
Thanks Tim!!

Tunes itself

It is even more easy to explain.. when you tune queries using adaptive technics.. etc.. all this can happen with the databases running.. rebuild indexes online.. etc.. must of tune operations nowadays can be done online.. it's not so difficult to have an AI ( Artificial Intelligence ) engine collecting data and taking some decisions based on internal tests.. statistics.. etc.. so.. this new level of automation/autonomous is just what we have but being done automatically.

Let's talk with what frequency we do these tasks:

Applying of Patches: time to time.. It's not so frequent.. so.. there are not too much changes to appreciate in the time we employ in this task

Upgrades: even less.. it's very seldom when we apply upgrades to the databases.. so.. there are not too much changes to appreciate in the time we employ in this task

Tunes itself: tuning.. yes.. many companies.. consultants.. etc.. employ a lot of time in constant tuning to databases for many reasons.. it's very common that some objects.. some queries.. some code.. etc are added, removed.. changed in the database and of course this generation variations in the performance. This point could impact relatively high the cost of many enterprises, because tuning a database constantly imply constant high costs. Of course.. we have to see how good the engine can be done this. We have to remember that currently we have available "Tuning advisors" and they are not perfect.. sometimes we apply some recommendations and the performance of some executions.. etc.. become worst.. so.. the factor of "Tuning itself".. really.. thinking very honest.. we have to see how that will work. If the results are very good and convenient.. definitely this factor will reduce a lot of cost to enterprises and of course will affect in some way DBAs specialized or in charge of tuning.

So.. one important thing to think.. You guess Enterprises with extreme important databases will allow to run all these tuning.. etc.. by itself without someone that can be monitoring what the database is doing.. ?

We are going to be honest.. Enterprises with critical databases will not allow to give the 100% of confidence to an engine software to tune itself on demand without control or supervision of a human. Of course, there are some databases.. under some levels of importance.. circumstances that fit that model.. to be tuned 100% without supervision but that not the regular cases.

That self tuning will have sections.. factor.. policies.. etc.. and DBAs will choose what elements could be tune by itself.. and which of them will keep under control of DBAs. At the end.. always DBAs will be needed.. Now DBAs will have more facilities to do their jobs more easy.. but that fact of thinking to replace totally the humans is something difficult to happen in spite of the engine can take correct decisions.

Let's describe a simple example.. Oracle Data Guard has the option to FAILOVER automatically under some circumstances, but for some possible humans errors could happen a database can failover when the humans did not want it.. so.. most of enterprises they fear to use FSF ( Fast Start Failover ), It's a mechanism that works perfect technically but to give the total control to the software to apply something critical to the data is something most of enterprises are not allow to implant.

In a general way I think autonomous databases will reduce tons of working hours for DBAs but will not represent in a practical and real way the elimination of Human presence.

So, autonomous databases promises the following:

Less time on Administration
- Less time on infrastructure
- Less time on patching, upgrades
- Less time on ensuring availability
- Less time on tuning

More time on Innovation
- More time on database design
- More time on data analytics
- More time on data polices
- and very important.. More time on securing data

So.. with databases on Cloud, DBAs must strength their skills in Security

Now Let's jump to another point of analysis..

17 years ago.. it was so simple to design a database architecture. You had to decide to lodge your database in a:
  • Server..
  • Or in a Mainframe
  • Or even in a regular desktop machine in some cases..
Now, when DBAs will decide where to lodge databases must think in many options.. such as:
  • - Servers..
  • - Virtual Machines
  • - Engineering Systems such as “Exadata”
  • - and more..
They have to decide if the architecture will be implanted:
  • - “On-Prem” as the regular option
  • - Cloud in our own data center ( Private Cloud )
  • - Hybrid Cloud
  • - Public Cloud, now the public Cloud has more options with autonomous databases in Oracle18c.. etc

and more and more..

Before was so simple to design a database architecture.. now not.. Now we have a lot of options.. each one with pros and cons. Basically is a sea of options..

So.. the question is.. Who will be in charge of designing all this.. ?
The General Managers ? obviously not.. this will decide by Database Architects.. ( DBAs )

Now DBAs will be asked more to understand the business rather than just the mechanics of keeping the database healthy and running

Before the cost of Servers.. services.. etc.. was not the main thing with what DBAs had to deal.. they were focused just in keeping the database running and implanting them. Now is different.. , now creating a database in Cloud, for example.. implies a decision of what kind of service to use and in what mode.. “Metered.. Non-Metered..” etc.. those factors impact directly the economy of the enterprise and the mode of operating these databases, so.. now the DBAs are more related into the business.

Some years ago from the first minute Oracle Corp. announced the release of “Oracle Cloud”, another level of expertise for DBAs was born, turning this business more complex..
DBAs will not be fired.. now the roles of DBAs are even more important.. the role of DBAs is simply shifting to be more of a role of architect..

Who must worry because of all these changes.. ?

Those kind of operational DBAs who are or were in charge just of doing simple tasks.. such as:
- Checking Backups..
- Checking storage..
- Making some reports..
- Installing Patchs..
- Installing softwares..
- Creating constantly new environments for developers..

All those easy and simple tasks.. have been removing with the addition of automation in Oracle Releases..

For instance, now at the time of creating an Oracle Database Cloud Service, there are many elements created automatically including:

- The creation of compute node ( Host )
- Installation of Software
- and more..

So.. with just this simple example.. you can realise that at the time of working with Oracle Database Cloud Service you do not have to install softwares anymore..

If you want to create a RAC configuration in a using Oracle Database Cloud Service is just some simple clicks and you could have complex RAC-DG configurations.

The era of regular and boring tasks is getting away.. so.. if you are a DBA in charge of those kind of simple tasks.. yes.. worry about your future..

Contrary of this, if you are a DBA in charge of implanting MAA architectures.. Cloud Databases, Exadata.. etc.. happy for you.. the importance of your role now will be bigger..

So.. now we come back to the central point of this article..
you can imagine what will the vision of a person that is starting as DBA with all these numerous options in the market.. A person that is barely learning SQL.. and “Create table..” etc.. when the market is talking about autonomous databases.. It’s a comparison like between a kid bike and one 1000cc Motorbike.. something like that..

My though about all this is:

#1.- DBAs with kind of regular maintenance tasks.. could be removing from their role or job if they do not evolve to be focused in all this new generation related to Cloud

#2.- DBAs that have got the privilege to work with Oracle Databases during the golden era of “On-prem” databases, now they are in the right time to grow in parallel with something that is defining the “before” and “after” of IT. Now the DBAs will have the opportunity to move into a more strategic role in their organisations and based on that the income "salary.. commissions.. etc" will be higher.. now you will be involved in important decisions for the company 

#3.- Sorry for new DBAs entering in this database world.. the journey you have to run is even harder than before to get the top level in this area because day by day are generated more and more options


1.- If you are a DBA kind #1 according to the previous description, perhaps you have several options.. evolve to #2, or if you decide to stay in the same role, your payment could be less over time or your role could be slowly disappear.

2.- If you are a DBA kind #2, keep on the track.. and your role will be more and more important in this field, by consequence your payment will have high probabilities to be increased. One more thing.. most of DBAs will not reach to be Cloud architects because this require time of study.. researching.. testing.. etc.. so.. feel proud if you reach it and take advantage of that inside your current or future organizations

3.-   If you are a DBA kind #3, do not give up.. do not see the journey too long as It really is, just keep learning and you will get the level to be a good competitor in the market.

No comments:

Post a Comment